hapsampler: Haplotype Sampler
In geo047/HapSampler: QTL-marker haplotype builder

Description Usage Arguments Details Value Examples

Marker-QTL haplotypes are sampled, and their probabilities calculated via MCMC methods.

1 2	hapsampler(data = NULL, trait = NULL, nchains = 3, runlength = 30, probthresh = 0.95)

`data`	a data frame object generated by running `read_data`.
`trait`	the column name in the data file of the trait. This must be specified.
`nchains`	the number of parallel chains to be run.
`runlength`	the run length of each chain
`probthresh`	a probability theshold value. Data on animals whose haplotype probability is less than `probthresh` will be ignored.

This function implements a MCMC algorithm for sampling the space of marker-QTL genotypes, conditional on the observed marker haplotypes for each animal, the trait values, and a penetrance function (i.e. the probability of the trait given the putative QTL alleles).

To guard against convergence issues, parallel chains are run by setting the argument nchains to some integer value greater than 1. runlength is the run length of a chain.

There are several plotting functions (such as plotLike, plotEffects, and plotScatter) available for viewing the results.

A list like object of class HS is returned with elements

trait: the trait name
nchains: the number of chains
canonical.hap: this is the haplotype that is the most common in the input data.
haplotypes: a data.frame object containing
- Haplotype: the haplotype index
- Count: the frequency of the haplotype in the input data file after data on those animals with a haplotype probability less than probthresh have been removed.
- MeanProb(Q): the probability of this haplotype carrying the Q qtl allele, averaged over the mean of the Q allele probabilities from each chain
- Prob(Q):Chain1: the mean probability of this haplotype carrying the Q qtl allele, for chain 1. This field is repeated for each chain that is run.

    complete.name <- system.file("extdata", "dataexample.dat", package="HapSampler")
    # read in phenotypic data which is space separated
    dt <- read_data(path=dirname(complete.name),
                                 file=basename(complete.name))
    
    #  perform analysis for the NAVEL trait
    hapres <- hapsampler(data=dt, trait="NAVEL", nchains=2, runlength=5)
    
    # a plot of the log likelihood for each chain against run length
    plotLike(hapres)